Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
mvectors
Browse files- pages/Semantic_Search.py +0 -20
- semantic_search/all_search_execute.py +0 -11
- utilities/mvectors.py +0 -2
pages/Semantic_Search.py
CHANGED
|
@@ -801,42 +801,22 @@ def render_answer(answer,index):
|
|
| 801 |
final_desc_ = "<p></p><p>"
|
| 802 |
for word_ in desc___:
|
| 803 |
str_=re.sub('[^A-Za-z0-9]+', '', word_).lower()
|
| 804 |
-
###### stemming and highlighting
|
| 805 |
-
|
| 806 |
-
# ans_text = ans['desc']
|
| 807 |
-
# query_text = st.session_state.input_text
|
| 808 |
-
|
| 809 |
stemmed_word = next(iter(set(stem_(str_))))
|
| 810 |
-
# print("stemmed_word-------------------")
|
| 811 |
-
# print(stemmed_word)
|
| 812 |
-
|
| 813 |
-
|
| 814 |
-
# common = ans_text_stemmed.intersection( query_text_stemmed)
|
| 815 |
-
# #unique = set(document_1_words).symmetric_difference( )
|
| 816 |
-
|
| 817 |
-
# desc__stemmed = stem_(desc__)
|
| 818 |
-
|
| 819 |
-
#print(str_)
|
| 820 |
if(stemmed_word in res___ or str_ in res___):
|
| 821 |
if(stemmed_word in res___):
|
| 822 |
mod_word = stemmed_word
|
| 823 |
else:
|
| 824 |
mod_word = str_
|
| 825 |
-
#print(str_)
|
| 826 |
if(res___.index(mod_word)==0):
|
| 827 |
-
#print(str_)
|
| 828 |
final_desc_ += "<span style='color:#ffffff;background-color:#8B0001;font-weight:bold'>"+word_+"</span> "
|
| 829 |
elif(res___.index(mod_word)==1):
|
| 830 |
-
#print(str_)
|
| 831 |
final_desc_ += "<span style='color:#ffffff;background-color:#C34632;font-weight:bold'>"+word_+"</span> "
|
| 832 |
else:
|
| 833 |
-
#print(str_)
|
| 834 |
final_desc_ += "<span style='color:#ffffff;background-color:#E97452;font-weight:bold'>"+word_+"</span> "
|
| 835 |
else:
|
| 836 |
final_desc_ += word_ + " "
|
| 837 |
|
| 838 |
final_desc_ += "</p><br>"
|
| 839 |
-
#print(final_desc_)
|
| 840 |
st.markdown(final_desc_,unsafe_allow_html = True)
|
| 841 |
elif("highlight" in ans and 'Keyword Search' in st.session_state.input_searchType):
|
| 842 |
test_strs = ans["highlight"]
|
|
|
|
| 801 |
final_desc_ = "<p></p><p>"
|
| 802 |
for word_ in desc___:
|
| 803 |
str_=re.sub('[^A-Za-z0-9]+', '', word_).lower()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 804 |
stemmed_word = next(iter(set(stem_(str_))))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 805 |
if(stemmed_word in res___ or str_ in res___):
|
| 806 |
if(stemmed_word in res___):
|
| 807 |
mod_word = stemmed_word
|
| 808 |
else:
|
| 809 |
mod_word = str_
|
|
|
|
| 810 |
if(res___.index(mod_word)==0):
|
|
|
|
| 811 |
final_desc_ += "<span style='color:#ffffff;background-color:#8B0001;font-weight:bold'>"+word_+"</span> "
|
| 812 |
elif(res___.index(mod_word)==1):
|
|
|
|
| 813 |
final_desc_ += "<span style='color:#ffffff;background-color:#C34632;font-weight:bold'>"+word_+"</span> "
|
| 814 |
else:
|
|
|
|
| 815 |
final_desc_ += "<span style='color:#ffffff;background-color:#E97452;font-weight:bold'>"+word_+"</span> "
|
| 816 |
else:
|
| 817 |
final_desc_ += word_ + " "
|
| 818 |
|
| 819 |
final_desc_ += "</p><br>"
|
|
|
|
| 820 |
st.markdown(final_desc_,unsafe_allow_html = True)
|
| 821 |
elif("highlight" in ans and 'Keyword Search' in st.session_state.input_searchType):
|
| 822 |
test_strs = ans["highlight"]
|
semantic_search/all_search_execute.py
CHANGED
|
@@ -274,10 +274,6 @@ def handler(input_,session_id):
|
|
| 274 |
vector_payload['neural']['product_description_vector']['filter']["bool"]["must"].append({"term": {"gender_affinity": st.session_state.input_gender}})
|
| 275 |
if(st.session_state.input_price!=(0,0)):
|
| 276 |
vector_payload['neural']['product_description_vector']['filter']["bool"]["must"].append({"range": {"price": {"gte": st.session_state.input_price[0],"lte": st.session_state.input_price[1] }}})
|
| 277 |
-
|
| 278 |
-
# print("vector_payload**************")
|
| 279 |
-
# print(vector_payload)
|
| 280 |
-
|
| 281 |
###### end of efficient filter applying #####
|
| 282 |
|
| 283 |
hybrid_payload["query"]["hybrid"]["queries"].append(vector_payload)
|
|
@@ -310,7 +306,6 @@ def handler(input_,session_id):
|
|
| 310 |
multimodal_payload['neural']['product_multimodal_vector']['filter'] = filter_['filter']
|
| 311 |
|
| 312 |
if(st.session_state.input_manual_filter == "True"):
|
| 313 |
-
print("presence of filters------------")
|
| 314 |
multimodal_payload['neural']['product_multimodal_vector']['filter'] = {"bool":{"must":[]}}
|
| 315 |
if(st.session_state.input_category!=None):
|
| 316 |
multimodal_payload['neural']['product_multimodal_vector']['filter']["bool"]["must"].append({"term": {"category": st.session_state.input_category}})
|
|
@@ -409,14 +404,9 @@ def handler(input_,session_id):
|
|
| 409 |
path = "retail-search-colbert-description/_search"
|
| 410 |
url = host + path
|
| 411 |
r = requests.get(url, auth=awsauth, json=hybrid_payload, headers=headers)
|
| 412 |
-
print(r.status_code)
|
| 413 |
-
#print(r.text)
|
| 414 |
response_ = json.loads(r.text)
|
| 415 |
-
print("-------------colbert ---- 3-----------")
|
| 416 |
-
#print(response_)
|
| 417 |
docs = response_['hits']['hits']
|
| 418 |
docs = cb.search(docs)
|
| 419 |
-
print("-------------COLBERT------------5------------------------------------------")
|
| 420 |
else:
|
| 421 |
single_query = hybrid_payload["query"]["hybrid"]["queries"][0]
|
| 422 |
del hybrid_payload["query"]["hybrid"]
|
|
@@ -525,7 +515,6 @@ def handler(input_,session_id):
|
|
| 525 |
arr.append(res_)
|
| 526 |
dup.append(doc['_source']['image_url'])
|
| 527 |
|
| 528 |
-
#print(arr)
|
| 529 |
return arr[0:k_]
|
| 530 |
|
| 531 |
|
|
|
|
| 274 |
vector_payload['neural']['product_description_vector']['filter']["bool"]["must"].append({"term": {"gender_affinity": st.session_state.input_gender}})
|
| 275 |
if(st.session_state.input_price!=(0,0)):
|
| 276 |
vector_payload['neural']['product_description_vector']['filter']["bool"]["must"].append({"range": {"price": {"gte": st.session_state.input_price[0],"lte": st.session_state.input_price[1] }}})
|
|
|
|
|
|
|
|
|
|
|
|
|
| 277 |
###### end of efficient filter applying #####
|
| 278 |
|
| 279 |
hybrid_payload["query"]["hybrid"]["queries"].append(vector_payload)
|
|
|
|
| 306 |
multimodal_payload['neural']['product_multimodal_vector']['filter'] = filter_['filter']
|
| 307 |
|
| 308 |
if(st.session_state.input_manual_filter == "True"):
|
|
|
|
| 309 |
multimodal_payload['neural']['product_multimodal_vector']['filter'] = {"bool":{"must":[]}}
|
| 310 |
if(st.session_state.input_category!=None):
|
| 311 |
multimodal_payload['neural']['product_multimodal_vector']['filter']["bool"]["must"].append({"term": {"category": st.session_state.input_category}})
|
|
|
|
| 404 |
path = "retail-search-colbert-description/_search"
|
| 405 |
url = host + path
|
| 406 |
r = requests.get(url, auth=awsauth, json=hybrid_payload, headers=headers)
|
|
|
|
|
|
|
| 407 |
response_ = json.loads(r.text)
|
|
|
|
|
|
|
| 408 |
docs = response_['hits']['hits']
|
| 409 |
docs = cb.search(docs)
|
|
|
|
| 410 |
else:
|
| 411 |
single_query = hybrid_payload["query"]["hybrid"]["queries"][0]
|
| 412 |
del hybrid_payload["query"]["hybrid"]
|
|
|
|
| 515 |
arr.append(res_)
|
| 516 |
dup.append(doc['_source']['image_url'])
|
| 517 |
|
|
|
|
| 518 |
return arr[0:k_]
|
| 519 |
|
| 520 |
|
utilities/mvectors.py
CHANGED
|
@@ -70,8 +70,6 @@ def search(hits):
|
|
| 70 |
token = tokens[index]
|
| 71 |
if(token!='[SEP]' and token!='[CLS]'):
|
| 72 |
query_token_vector = np.array(i)
|
| 73 |
-
print("query token: "+token)
|
| 74 |
-
print("-----------------")
|
| 75 |
scores = []
|
| 76 |
for m in with_s:
|
| 77 |
m_arr = m.split("-")
|
|
|
|
| 70 |
token = tokens[index]
|
| 71 |
if(token!='[SEP]' and token!='[CLS]'):
|
| 72 |
query_token_vector = np.array(i)
|
|
|
|
|
|
|
| 73 |
scores = []
|
| 74 |
for m in with_s:
|
| 75 |
m_arr = m.split("-")
|